CN111208169A

CN111208169A - Pulse eddy current thermal imaging defect detection method suitable for natural cracks

Info

Publication number: CN111208169A
Application number: CN202010104000.1A
Authority: CN
Inventors: 殷鹰; 陈娟; 高斌; 刘泽伟; 邓兵; 何媛媛; 吴迎春; 王莉
Original assignee: SICHUAN SPECIAL EQUIPMENT TESTING RESEARCH INSTITUTE; University of Electronic Science and Technology of China
Current assignee: SICHUAN SPECIAL EQUIPMENT TESTING RESEARCH INSTITUTE; University of Electronic Science and Technology of China
Priority date: 2020-02-20
Filing date: 2020-02-20
Publication date: 2020-05-29

Abstract

The invention discloses a pulse eddy current thermal imaging defect detection method suitable for natural cracks, which starts with the improvement of a sensing framework and takes a newly defined evaluation index, namely thermal contrast, as the maximum principle to determine the lifting distance and the horizontal distance of a straight wire so as to determine an interval eddy current thermal imaging area for the characteristics of a complex surface strong noise signal, discontinuity of the natural cracks and internal adhesion. The thermal contrast in the invention considers the temperature change of the pixel points of the target area and the background area, also considers the average temperature of the target area and the background area, considers the eddy current density and the uniformity of the eddy current field, is very useful for detecting the complex natural cracks of the background noise signal, and improves the detection efficiency.

Description

Pulse eddy current thermal imaging defect detection method suitable for natural cracks

Technical Field

The invention belongs to the technical field of defect detection, and particularly relates to a pulse eddy current thermal imaging defect detection method suitable for natural cracks.

Background

In the industries of automobiles, nuclear energy, aviation, railways, shipbuilding and the like, equipment and parts can generate cracks during manufacturing and using. These cracks are non-damage-causing and are referred to as natural cracks, for example, unfused, unwelded, fatigue-forming cracks in pressure vessel welds and stress corrosion-forming cracks, as well as rolling contact fatigue-forming cracks (scale-like peeling cracks and diagonal cracks) in steel rails. Natural cracks have great potential safety hazards, and cracks in welding seams can influence the compactness of the pressure container, so that major accidents are caused; after the cracks formed by fatigue on the rails have propagated to a certain extent, serious derailment accidents may occur during the operation of the train. Therefore, analysis and detection of natural cracks is particularly important.

The defect detection of pulse eddy current thermal imaging belongs to active infrared thermal imaging defect detection, and the method uses high-frequency alternating current to generate eddy current on the surface of a detected test piece, and according to the Joule law, part of the eddy current can be converted into Joule heat which is propagated on the surface and the inside of the test piece, and the temperature on the surface of the test piece can be changed. When the test piece has defects, the eddy current can detour, the joule heat distribution can also generate abnormity, and the temperature abnormity can be detected through the thermal infrared imager, so that the purpose of defect detection is achieved.

The pulse eddy current thermal imaging defect detection can capture the abnormality of an eddy current field and a thermal field caused by the defects on the tested piece, and then judge the position and the size of the defects. The detection sensitivity is related to the distribution of the magnetic field and the eddy current field generated by the inductor, for example, the excitation range of the magnetic field and the eddy current field determines the single detectable area of the detection system, and the density of the eddy current generated on the surface of the test piece affects the detection capability and the like.

The test pieces such as welding seams, steel rails and the like belong to complex structures due to rough surfaces. Cracks (natural cracks) that occur during manufacturing and service due to alternating stress and corrosion are difficult to detect. The following difficulties exist in defect detection: on one hand, the complex surface can generate corresponding eddy current field change and temperature abnormity to interfere the detected defect signal, so that the detected defect signal is difficult to extract; on the other hand, the natural crack defect is adhered, and the conductivity of the natural crack defect is not strictly 0, so that the defect is not easy to detect. If the test piece can be uniformly heated, the defect signals can not be covered by the noise signals, and meanwhile, weak defect signals can also be effectively extracted. Therefore, the requirement of natural crack detection on heating uniformity is high.

There are two main ways to improve the heating uniformity: firstly, the sensing architecture is improved, and secondly, excitation parameters are optimized. The scholars have made major research progress in these two areas.

And (4) improvement of a sensing architecture. Common sensing architectures include straight wire coils, spiral coils, helmholtz coils, magnetic core-surrounding coils, and the like.

The straight lead coil forms a high-temperature zone on the surface of the test piece, the eddy current of the high-temperature zone is uniform, the eddy current field in the direction vertical to the coil can be weakened, and the straight lead coil is often used in the pulse eddy current thermal imaging defect detection. Bin Gao used a straight wire coil in fatigue evaluation of the turbine; the Xiaoqing Li uses a straight wire for quantitative detection of surface cracks, and researches the influence of the sizes of the cracks on detection results.

The spiral coil generates a radial magnetic field and a circular eddy current field on the surface of the test piece, and can be used for detecting defects at different angles. Benmoussat uses a spiral coil to detect surface defects of a metal material.

Helmholtz coil can produce even magnetic field in two coil middle areas through setting up suitable coil radius and interval, is applicable to small-size test piece and detects. Oswald-Trnta detects the metal surface defects by using a Helmholtz coil, and evaluates defect depth information by using a phase difference; jianping Peng uses a helmholtz coil to detect rail rolling contact fatigue cracks.

The magnetic core is added into the spiral coil around the coil, so that more magnetic energy can be generated in the test piece. Compared with a single spiral coil, the eddy current field on the surface of the test piece is larger. The Jianan Zhao establishes a theoretical model of the magnetic core ring structure, explains the advantages of the magnetic core ring structure in electromagnetic thermal imaging detection, and verifies the characteristics of large detection range, capability of detecting cracks at various angles and the like by simulation and experiments.

Excitation parameter optimization. Lahiri detects hidden defects using low frequency excitation [18 ]; H.Shen investigated the effect of lift-off, excitation current, excitation frequency coil turns on the historical temperature curve [19 ].

When the pulse eddy current thermal imaging defect detection is used for detecting the natural cracks of the complex structure, the signal to noise ratio of the defect detection result is lower because the surface noise signal of the complex structure is large, and the discontinuity and internal adhesion phenomena exist in the natural cracks. However, the natural cracks of pressure vessels, steel rails and the like are harmful, and if the natural cracks cannot be detected accurately and effectively in time, economic loss can be caused, and casualties can also be caused. Therefore, how to further improve the detection rate of the natural cracks by optimizing the excitation parameters and the sensing framework is an important consideration in the detection.

Disclosure of Invention

The invention provides a pulse eddy current thermal imaging defect detection method suitable for natural cracks aiming at the characteristics of strong noise signals on complex surfaces, discontinuity of natural cracks and internal adhesion, and excites more uniform eddy current fields and temperature fields while considering eddy current density at the defects (natural cracks) so as to improve detection effectiveness and efficiency.

In order to achieve the above object, the present invention provides a method for detecting defects of natural cracks by pulsed eddy current thermal imaging, comprising the steps of:

(1) adopting a straight wire as an excitation coil, wherein the straight wire is parallel to the surface of the test piece, the defect (natural crack) is positioned on the test piece on one side of the vertical lower part of the straight wire, the vertical distance (lift-off distance) from the center of the straight wire to the surface of the test piece is h, and the horizontal distance from the vertical lower part of the center of the straight wire to one end, close to the defect (natural crack), of the defect (natural crack) is m;

(2) setting a defect-containing area as a target area, setting a non-defect area with the same area as the area adjacent to one side of the target area along the straight lead as a background area, combining the target area and the background area together to form a scene area, wherein the area formed by the scene area extending along the front and back of the straight lead is an interval eddy current thermal imaging area;

(3) defining an evaluation index: thermal contrast Δ T

Wherein N is_T、N_B、N_SNumber of pixels respectively representing target area, background area and scene area, N_S＝N_T+N_B，σ_T ²、σ_B ²Variance, mu, respectively representing temperature (pixel value) in the target region and in the background region_T、μ_BMean values representing temperatures (pixel values) in the target region and in the background region, respectively;

(4) inputting excitation current (high-frequency alternating current) to the straight conductor, carrying out temperature acquisition on the interval eddy current thermal imaging area by using a thermal infrared imager to obtain a thermal image, and then determining the lifting distance and the horizontal distance of the straight conductor according to the principle of maximum thermal contrast delta T to obtain a final defect detection thermal image.

The invention aims to realize the following steps:

the invention is suitable for the detection method of the pulse eddy current thermal imaging defect of the natural crack, starts from improving the sensing framework for the characteristics of a complex surface strong noise signal, discontinuity of the natural crack and internal adhesion, and determines the lifting distance and the horizontal distance of the straight wire by taking a newly defined evaluation index, namely thermal contrast, as the maximum principle, thereby determining the interval eddy current thermal imaging area. The thermal contrast in the invention considers the temperature change of the pixel points of the target area and the background area, also considers the average temperature of the target area and the background area, considers the eddy current density and the uniformity of the eddy current field, is very useful for detecting the complex natural cracks of the background noise signal, and improves the detection efficiency.

Drawings

FIG. 1 is a flow chart of one embodiment of a method for detecting defects by pulsed eddy current thermal imaging for natural cracks according to the present invention;

FIG. 2 is a schematic diagram of a sensing architecture setup;

FIG. 3 is a top view of the sensing architecture arrangement of FIG. 2;

FIG. 4 is a graph of the eddy current distribution on the surface of the test piece;

FIG. 5 is a graph showing the distribution of eddy current density on the surface of a test piece at different lift-off distances;

FIG. 6 is a schematic view of three regions of the test piece surface featuring different eddy current density distributions;

FIG. 7 is a schematic diagram of eddy current detour at a defect, (a) straight wire excitation, (b) ideal uniform excitation;

FIG. 8 is a schematic diagram of a sensing architecture setup for different lift-off distances;

FIG. 9 is a schematic diagram of a sensing architecture setup for different horizontal distances;

fig. 10 is a isotherm diagram of surface defect detection at different lift-off distances, wherein (a) h is 1+3.5mm, b) h is 2+3.5mm, (c) h is 3+3.5mm, (d) h is 4+3.5mm, (e) h is 5+3.5mm, (f) h is 10+3.5mm, and (g) h is 20+3.5 mm;

FIG. 11 is a graph of surface natural crack detection data for different lift-off distances;

fig. 12 is a temperature contour plot of different levels of distance surface defect detection, wherein (a) m is 2mm, (b) m is 3mm, (c) m is 3.5mm, (d) m is 4mm, (e) m is 4.5mm, (f) m is 5mm, (g) m is 6 mm;

FIG. 13 is a graph of surface natural crack detection data for different horizontal distances;

fig. 14 is a isotherm diagram of subsurface defect detection at different lift-off distances, where (a) h is 1+3.5mm, (b) h is 2+3.5mm, (c) h is 3+3.5mm, (d) h is 4+3.5mm, (e) h is 5+3.5mm, (f) h is 10+3.5mm, (g) h is 4+3.5 mm;

fig. 15 is a thermogram of different horizontal distance subsurface defect detection, (a) m 2mm, (b) m 3mm, (c) m 3.5mm, (d) m 4mm, (e) m 4.5mm, (f) m 5mm, (g) m 6 mm;

FIG. 16 is a graph of subsurface natural crack detection data for different lift-off distances;

FIG. 17 is a graph of subsurface natural crack detection data for different horizontal distances;

FIG. 18 is a diagram of an experimental setup;

FIG. 19 is a model view of a weld;

FIG. 20 is a schematic view of a defect in a weld;

FIG. 21 is a permeation test result of weld defects;

fig. 22 is a real photograph and a magnified result of defect 1, in which (a) is an original photograph and (b) is a 60-fold magnified result;

fig. 23 shows the results of experiments with different lift-off distances for defect 1 and defect 2, where (a) h is 2mm, (b) h is 4mm, (c) h is 6mm, (d) h is 7mm, (e) h is 8mm, and (f) h is 10 mm;

fig. 24 shows the results of experiments with different lift-off distances for defect 3, where (a) h is 2mm, (b) h is 4mm, (c) h is 6mm, and (d) h is 9 mm;

fig. 25 shows the results of experiments at different horizontal distances of defect 1, where (a) m is 2.5mm, (b) m is 3.5mm, (c) m is 4.5mm, (d) m is 5.5mm, (e) m is 6.5 mm;

fig. 26 shows the results of experiments with different horizontal distances of defect 4, where (a) m is 3.5mm, (b) m is 4.5mm, (c) m is 5.5mm, and (d) m is 6.5 mm.

Detailed Description

The following description of the embodiments of the present invention is provided in order to better understand the present invention for those skilled in the art with reference to the accompanying drawings. It is to be expressly noted that in the following description, a detailed description of known functions and designs will be omitted when it may obscure the subject matter of the present invention.

FIG. 1 is a flow chart of an embodiment of a method for detecting defects by pulsed eddy current thermal imaging, which is suitable for natural cracks.

In this embodiment, as shown in fig. 1, the method for detecting defects by pulsed eddy current thermal imaging suitable for natural cracks according to the present invention includes the following steps:

step S1: setting sensing architecture

A straight wire is used as an excitation coil, the straight wire is parallel to the surface of a test piece, a defect (natural crack) is located on the test piece on one side of the vertical lower side of the straight wire, the vertical distance (lift-off distance) from the center of the straight wire to the surface of the test piece is h, and the horizontal distance from the vertical lower side of the center of the straight wire to one end, close to the defect (natural crack), of the defect (natural crack) is m.

In the present embodiment, the sensing structure is set up as shown in fig. 2 and 3, wherein the region a is a target region containing a defect, and the area thereof is 2 × 1mm²And the area b is a non-defect area which is adjacent to the area a and has the same area, namely a background area. In the present embodiment, the natural crack has a shape and a position as shown in fig. 2 and 3, and the natural crack has a vertical direction to the coil (excitation coil, i.e. straight wire), because the disturbance generated by the vertical natural crack is large and is most easily detected. The vertical distance (lift-off distance) from the center of the straight wire to the surface of the test piece is h, and the horizontal distance from the vertical lower part of the center of the straight wire to one end of the defect (natural crack) close to the straight wire is m. w and l represent the length and width of the natural crack, respectively.

Step S2: selecting an interval eddy current thermal imaging region

The defect-containing region is set as a target region, a non-defect region with the same area as the target region adjacent to one side of the straight conducting wire is set as a background region, the target region and the background region are combined to form a scene region, and a region formed by the scene region extending along the front and back of the straight conducting wire is a region eddy current thermal imaging region.

Johannes Vrana researches the distribution of eddy currents on the surface of a tested piece aiming at a simple model that an exciting coil is a straight wire and the tested piece is large enough (can be approximated as a semi-infinite body). The results show that the values of the eddy current density on both sides of the straight wire in the direction perpendicular to the straight wire (as an excitation coil) exhibit a lorentz curve, with the maximum value of the induced eddy current being directly below the straight wire, as shown in fig. 4. As can be seen from fig. 4, as the distance h from a certain point on the surface of the test piece to the center of the straight wire increases, the eddy current density at the point gradually decreases, and the eddy current density changes from a sharp decrease to a slow decrease. That is, in the course of the distance increase, the value of the eddy current density is decreasing, and the rate of decrease thereof becomes gradually slower. The effect that causes the eddy currents to be so distributed is known as the proximity effect.

For a test piece of limited width, the eddy current density will rise again at the edges due to the edge effect. The fringe effect means that an alternating current generates an alternating magnetic field, and an electric field in a direction opposite to the current is induced to block the flow of the current according to the law of electromagnetic induction, so that the resistance of the middle current of the alternating current field is large, and the resistance of the fringe current is small, so that the middle density of the alternating current field is small and the fringe density is large. The invention aims at the distribution of the eddy current density near the straight wire, so the eddy current distribution condition of the whole test piece is not drawn.

On the other hand, under certain experimental conditions, test piece surface eddy current density distribution curves of different lift-off distances (eddy current density distribution curves of 1mm, 2mm, 3mm, 4mm, 5mm, 10mm and 20mm in sequence from top to bottom) were plotted, as shown in fig. 5. The lift-off distance indicated in fig. 5 is the vertical distance from the center of the straight wire to the surface of the test piece. It can be seen that, as the lift-off distance increases, the peak value of the eddy current density induced on the surface of the test piece decreases, and the decrease of the eddy current density from the middle to the two sides gradually becomes gentle, i.e. the uniformity is enhanced.

Taking a point P on the surface of the test piece, fig. 4 illustrates that the eddy current density value decreases and the uniformity increases as the point P is located from near to far from the horizontal distance of the straight wire. Fig. 5 illustrates that as point P is moved from near to far from the vertical lift-off of a straight wire, the eddy current density value at point P is also decreasing and the uniformity is increasing. The horizontal distance and the lifting distance between the point P and the straight wire are changed, and the eddy current distribution characteristics of the point P have similarity.

The above analysis shows that there are three regions on the surface of the test piece with different eddy current distribution characteristics, as shown in fig. 6. The eddy current density in the area A is high but the uniformity is poor, the eddy current density in the area B is enough and the uniformity is good, and the eddy current density in the area C is uniform but the value is too small to generate enough heat to detect defects. The size of each interval is influenced by the vertical lift-off distance between the straight conducting wire and the test piece. The B area, namely an area formed by extending the scene area back and forth along the straight wire is taken as an interval eddy current thermal imaging area.

Step S3: definition of evaluation index, i.e., thermal contrast Δ T

Wherein N is_T、N_B、N_SNumber of pixels respectively representing target area, background area and scene area, N_S＝N_T+N_B，σ_T ²、σ_B ²Variance, mu, respectively representing temperature (pixel value) in the target region and in the background region_T、μ_BRespectively, mean values of temperatures (pixel values) in the target region and in the background region.

In the existing thermal contrast, the area weighted average temperature difference is calculated from the average temperature of the target and background areas, and the temperature change of pixel points of the target and background areas is ignored, which is equivalent to sigma_T ²＝σ _B ²0; the root mean square temperature difference considers the temperature of the pixel point in the target area, but ignores the temperature characteristic of the background and the average temperature of the target area, namely A/mm²。

In order to solve the problems existing in the traditional thermal contrast definition, the invention provides a new thermal contrast Δ T definition method. Compared with the traditional definition of the thermal contrast, the thermal contrast delta T of the invention comprises the pixel number of the target area and the background area, and the mean value and the variance of the temperature value of each pixel. It is defined as follows:

ΔT＝σ_T|S-σ_B|S(1)

wherein sigma_T|SAnd σ_B|SRespectively representing the target pixel temperature standard deviation and the background pixel temperature standard deviation at the average scene temperature. The mean and variance of the temperatures of the target area, background area and scene area are defined as follows:

wherein, P_i ^T,P_i ^B,P_i ^STemperature values, N, of pixel points respectively representing a target region, a background region and a scene_T, N_B,N_SRespectively representing the number of pixel points. Wherein N is_S＝N_T+N_B。σ_T|S ²And σ_B|S ²The calculation can be as follows:

Δ T in the formula (1) can be re-expressed as

The new definition of the thermal contrast delta T considers the temperature change of the pixel points of the target area and the background area and also considers the average temperature of the target area and the background area, the eddy current density and the uniformity of the eddy current field are considered, the method is very useful for detecting the complex natural cracks of the background noise signals, and the detection efficiency is improved.

Step S4: determining the lift-off distance and the horizontal distance to obtain a final defect detection thermal image

Inputting excitation current (high-frequency alternating current) to the straight conductor, carrying out temperature acquisition on the interval eddy current thermal imaging area by using a thermal infrared imager to obtain a thermal image, and then determining the lifting distance and the horizontal distance of the straight conductor according to the principle of maximum thermal contrast delta T to obtain a final defect detection thermal image.

In pulsed eddy current thermal imaging, if there is a defect on the surface of the test piece, the eddy current may detour around the defect due to changes in electrical conductivity and magnetic permeability. Fig. 7 compares the eddy current detour in both the case of a straight wire (excitation coil or coil for short) excitation and the case of an ideal uniform excitation. The line with horizontal arrows represents the eddy current, and the thickness thereof represents the magnitude of the eddy current density. In fig. 7(a), the two ends of the defect have different distances from the straight conductor, and the eddy current density in the surrounding area has different values, so that the generated disturbance is different. In FIG. 7(b), the eddy current density values in the regions around the two ends of the defect are the same, and the generated disturbance is also in the same order of magnitude.

If the defect size is large, the disturbance caused by the defect is strong, and when the straight lead is adopted for excitation, even if the straight lead is close to the defect, the temperature of the defect is obviously abnormal, and the defect can be captured by a thermal imager.

However, for natural cracks (defects), because the size of the defect is small, the eddy current field disturbance is small, and therefore, when the natural cracks (defects) are excited by a straight wire, the influence caused by the nonuniformity of the eddy current density cannot be ignored. The eddy current density of the area around the side (short for the near side) of the defect closer to the straight wire is high, but the generated disturbance is not easy to be captured due to the violent change. And the eddy current density of the area around the side (far side for short) of the defect far away from the straight wire is small, but the distribution is uniform, and under the condition that the eddy current density value is enough to be detected, the disturbance can be captured, so that the defect is identified.

The strength of the eddy current and the uniformity of eddy current distribution jointly determine whether the defect can be detected, and for the detection of natural cracks, the strength of the eddy current needs to be ensured, and the distribution of the eddy current needs to be more uniform as much as possible. The uniform eddy current field produces a uniform thermal field that facilitates defect detection and evaluation. Especially, the method has important significance for extracting weak signals aiming at natural cracks.

According to the above analysis, selecting a region with sufficient eddy current density and uniform distribution can effectively reduce the influence of the non-uniformity of the eddy current boundary on the crack detection (especially, the natural crack), i.e., the region B in fig. 6. The above research also shows that the relative position with the defect has a non-negligible influence on the detection result, and when the defect position is known, a proper straight wire placement position can be selected directly according to the principle. When the defect position is unknown, the straight wire can be scanned on the surface of the test piece at a low speed, and the area where the straight wire passes necessarily covers the optimal detection area, namely the interval eddy current thermal imaging area.

First, simulation experiment

(1) Simulation experiment setup

In order to further study the influence of the horizontal and vertical distances from a certain point on the surface of the test piece to the center of the straight wire on the eddy current density of the point, COMSOL software is adopted to simulate the tiny defects on the surface and the subsurface. The exciting coil adopted by simulation is a straight wire, the radius of the straight wire is 3.5mm, and the length of the straight wire is 200 mm. The excitation frequency was 256kHz, the excitation current 300A, the initial temperature 293.15K, and the heating time 300 ms. The electromagnetic parameters of the test pieces are shown in table 1.

Parameter(s)	Value of
		Conductivity (S/m)	1.370e7
Relative magnetic permeability	1
		Density (kg/m3)	8030
Specific heat capacity (J/(kg K))	502
		Thermal conductivity (W/(m K))	12.1

TABLE 1

The excitation time is selected according to the following steps: the eddy current energy generated by the exciting source per unit volume per unit time is constant, and the temperature of the surface of the test piece is increased along with the excitation. Due to the thermal diffusion, the heat generated in the test piece gradually tends to stabilize at the end of the heating phase. If excitation is not stopped, the temperature on the test piece can slowly rise, the change rate of the temperature is low, and the improvement of the detection effect is not significant. The optimal heating time depends on the time for the eddy current and thermal diffusion to reach equilibrium, and is also related to the resolution of the camera and the heat resistance of the test piece. A smaller excitation time of 300ms was chosen in this simulation.

The excitation current is selected according to: the temperature rise of the surface of the test piece is in direct proportion to the size of the excitation current. The larger the excitation current, the more intense the heat diffusion. Then the thermal contrast, thermal mode, signal to noise ratio are enhanced at defect and non-defect without affecting the material, so in this simulation the excitation current should be chosen larger. In this simulation, the excitation current was set to 300A.

The natural crack size is small, typically on the order of microns, so the natural crack size is set in the simulation to be 1mm in length by 0.2mm in width by 0.3mm in depth.

According to the principle that the thermal contrast Delta T is maximum, the lifting distance and the horizontal distance of the straight conducting wire are determined as follows: fixing the lifting distance of the straight wire, changing the horizontal distance, determining the horizontal distance or fixing the horizontal distance of the straight wire when the thermal contrast delta T is maximum, changing the lifting distance, and determining the lifting distance when the thermal contrast delta T is maximum. By changing the lift-off distance h and the horizontal distance m, the situation of different vertical distances and different horizontal distances from the straight wire to the crack can be studied, as shown in fig. 8 and 9.

(2) Simulation experiment of different lift-off distances of surface defects

In order to study the effect of lift-off distance on the eddy current field and thermal field distribution, the conditions of different lift-off distances of the straight line and the surface defect (lift-off distances h of 1mm, 2mm, 3mm, 4mm, 5mm, 10mm, 20mm) were simulated, and the resulting isotherm diagram is shown in fig. 10.

In fig. 10, white dotted lines indicate straight line center lines, and the lift-off distance gradually increases from fig. 10(a) to fig. 10 (g). It is clearly seen that the maximum temperature in fig. 10 decreased from 293.43 deg.c (fig. 10(a)) to 293.15 deg.c (fig. 10(g)) as the lift-off distance increased. The temperature change in fig. 10(a) is also large and decreases rapidly from the straight wire to both sides, the isotherm diagram is divided into 9 layers according to the temperature value, the number of layers of the isotherm diagram in fig. 10(b) (c) is equivalent, the isotherm diagram is divided into 5 layers in fig. 10(e), only 3 layers in fig. 10(f) are present, and the temperatures in fig. 10(g) are equivalent. On the other hand, the hot spot at one end of the defect closer to the straight wire is gradually separated from the initial submerged state, the hot spot is almost completely submerged in fig. 10(a), and fig. 10(c) is gradually separated from the surrounding high temperature region until finally becoming a hot spot (fig. 10 (g)). In order to compare the detection effects of natural cracks under different lift-off distances, the mean value and the variance of the surface eddy current density of the defect-free test piece under each lift-off distance are calculated, wherein the mean value represents the strength of an eddy current field, and the variance represents the uniformity of the eddy current field. And the thermal contrast (Δ T newly proposed in the present invention) of the defective area and the non-defective area in the defective test piece was calculated, the values thereof are shown in table 2, and the changes at different lift-off distances are plotted in fig. 11.

Table 2 shows the surface natural crack detection data for different lift-off distances.

Distance of lift (mm)	Mean value of eddy currents (A/mm)²)	Variance of eddy current	ΔT
					1	16.8889	0.4898	0.0723
2	14.2165	0.3809	0.0759
				3	12.1884	0.3176	0.0766
4	10.5806	0.2849	0.0726
				5	9.2233	0.2563	0.0682
10	5.3728	0.1629	0.0464
				20	2.0622	0.0662	0.0187

TABLE 2

FIG. 11 is a graph of surface natural crack detection data for different lift-off distances.

As can be seen from Table 2 and FIG. 10, as the lift-off distance increases, the mean value of the eddy current density decreases, and the value is 16.8889A/mm²Down to 2.0622A/mm². And the density of the vortex drops faster at the stage of 1mm to 10mm and slows down after 10 mm. Overall, the eddy current field is reduced.

On the other hand, as the lift-off distance increases, the variance of the eddy current density is reduced from 0.4898 to 0.0662, and the overall size is reduced, namely, the uniformity of the eddy current field is enhanced.

During the increase of the lift-off distance, the thermal contrast of the defect and non-defect area rises from 0.0723 to 0.766 and then falls to 0.0187 again, i.e. Δ T increases and then decreases, reaching a peak at 3 mm.

The high eddy current strength and the uniform eddy current distribution are beneficial to defect detection, and the eddy current strength and the uniform eddy current distribution compete with each other in the process of increasing the lift-off distance. When the lift-off distance is increased from 1mm to 3mm, the value of the eddy current density is reduced, the eddy current distribution uniformity is enhanced, and the final thermal contrast is increased, which shows that the eddy current density uniformity becomes a main factor for determining the thermal contrast at the stage, so that the thermal contrast is increased, and the influence is larger than that of the eddy current strength. After lift-off greater than 3mm, the eddy current density decreases in value and uniformity is still increasing, and the final thermal contrast is decreasing, indicating that at this stage, strength becomes the dominant factor in determining thermal contrast, making thermal contrast lower, which affects uniformity more.

(3) Simulation experiment of different horizontal distances of surface defects

The eddy current density and the eddy current field uniformity affect the thermal contrast together, and because the changes of the eddy current density and the uniformity under different horizontal distances are not consistent, the conditions of different horizontal distances are simulated. From the analysis of fig. 5, it can be seen that if lift-off is too small, the eddy current drops too quickly and the effect of uniformity is easily negligible, so a larger lift-off should be chosen to balance the eddy current density and uniformity, so the fixed lift-off distance is 5mm here. Since the larger the excitation current, the larger the eddy current density value induced in the test piece, the excitation current was selected to be 900A. The resulting isotherm diagram is shown in fig. 12.

As can be seen from fig. 12, the lower end of the defect generates a high temperature region due to the disturbance, which is submerged by the heat generated from the straight wire at a horizontal distance of 2mm to 3.5 mm. At horizontal distances of 3mm and 4.5mm, the hot spot at the lower end of the defect is separated from the noise. When the horizontal distance is further increased, the induced eddy current density is small, so that sufficient disturbance cannot be generated at the defect, and a hot spot cannot be observed at the lower end of the defect.

The mean and variance of the eddy current density of the surface of the test piece without defects and the thermal contrast of the test piece with defects at different horizontal distances were calculated, the calculation results are shown in table 3, table 3 is the surface natural crack detection data at different horizontal distances, and the changes at different horizontal distances are plotted in fig. 13.

TABLE 3

From Table 3 and FIG. 13, it can be seen that the eddy current density value is 24.3034A/mm as the horizontal distance increases²Reduced to 16.4531A/mm²The decrease is faster when the horizontal distance is from 2mm to 5mm, and becomes slower after 5 mm. Vortex density in generalThe degree is reduced, and the eddy current field is weakened.

With the increase of the horizontal distance, within 3mm, the variance of the eddy current density is not greatly changed, in fact, because the horizontal distance is calculated from the central line of the straight wire, and the radius of the straight wire is 3.5mm, the current on the straight wire is also concentrated on the outer side of the straight wire according to the edge effect of the alternating current field. The 3mm is here the mapping of the area within the radius of the straight wire on the surface of the specimen. After 3mm, the variance of the vortex density is reduced from 2.2642 to 1.7500, and the uniformity is enhanced.

The resulting defect thermal contrast Δ T does not change monotonically during the increase in horizontal distance, similar to previous analysis of lift-off change, because eddy current density and uniformity work together, sometimes with the former playing a dominant role, causing the thermal contrast to increase, and sometimes with the latter playing a dominant role, causing the thermal contrast to decrease.

The above simulation results show that when the vertical lift-off distance and the horizontal distance between the straight wire and the natural crack are increased, the eddy current density is reduced, and the uniformity of eddy current distribution is enhanced. In this process, eddy current density and uniformity affect the detection result. Uniformity usually affects the inspection result (natural cracks have high requirement on uniformity due to small disturbance) first, and then eddy current density is insufficient to affect the inspection result. Therefore, the key to the inspection is to select a suitable region with a relatively high uniformity of the eddy current distribution and a sufficient eddy current density value. In the above simulation of surface defects, the following conclusions can be drawn: 1. when a smaller horizontal distance is fixed, the optimal vertical lift-off distance is 3 mm; 2. the optimum horizontal distance is 4mm when the fixed vertical lift-off distance is 5 mm.

(4) Simulation experiment of different lifting distances and different horizontal distances of subsurface defects

To investigate whether the above conclusions hold for subsurface defects, simulations were performed to hide subsurface defects with a depth of 0.01 mm. Isotherm plots for different lift-off distances and different horizontal distances are shown in fig. 14 and 15, respectively. The eddy current density mean and eddy current distribution variance for a defect-free specimen and the thermal contrast data for a defective specimen are plotted in fig. 16 and 17.

From the isotherm graph 14 and the detection data graph 16, it can be seen that as the vertical lift-off distance between the straight wire and the subsurface defect increases, the maximum temperature value of the excitation decreases, the range level of the temperature changes from more to less, and the same conclusion is obtained in the case of different lift-off of the surface defect. On the other hand, the value of the eddy current density decreases, and the uniformity of the eddy current distribution gradually increases and tends to be stable. When the lift-off distance is increased from 1mm to 10mm, the value of the eddy current density decreases,

the uniformity is enhanced and then stabilized, and the final thermal contrast is gradually high and gradually low and then slightly raised, which shows that the uniformity plays a main role; when the lift-off distance is larger than 10mm, the value of the eddy current density is reduced, the eddy current distribution uniformity is enhanced, and the final thermal contrast is reduced, which indicates that the eddy current density begins to influence the detection result.

As seen from the isotherm graph 15 and the inspection data graph 17, the eddy current density decreases as the horizontal distance increases, the uniformity of the eddy current distribution does not change monotonously, and the thermal contrast is highest at the horizontal distance of 3.5 mm. The peak point of thermal contrast is the minimum value in the eddy current density variance curve, which reflects that the more uniform region is more beneficial to defect detection under the condition of reduced eddy current density peak value and non-monotonic variation of uniformity.

The analysis shows that the subsurface defect also meets the interval eddy current theory, namely, the surface eddy current field of the defect-free test piece is weakened along with the increase of the vertical distance and the horizontal distance between the straight lead and the defect, and the eddy current distribution uniformity is integrally enhanced. The two influence the thermal contrast of a defect and a non-defect area in a test piece containing defects together, when the eddy current field intensity and the uniformity are poor, the change of the eddy current boundary is large, and a defect signal can be submerged by the change of the eddy current field; when the uniformity is good and the eddy current field is weak, the disturbance generated by the defect signal is too small to be separated from the thermal signal generated by the straight wire, and the defect can be detected with higher thermal contrast only when the eddy current field is strong enough and is more uniform. This phenomenon is particularly evident in natural crack detection, where due to the small size of natural cracks, both a large eddy current density is required to cause sufficient disturbance and an even eddy current field is required to reduce the influence of eddy current boundary inhomogeneity on its small disturbance.

Second, detection experiment

(1) Introduction of experimental System

In this example, the experimental apparatus is shown in fig. 18 and includes a power generator, water cooling, an induction heating device, a thermal infrared imager, and a PC. The thermal imager has a temperature resolution of 0.08K and a frame frequency of 200Hz, since the optimal time scale of the natural crack detection and quantification stage is 10^-4And s, selecting the maximum frame frequency of 200Hz to be close to the time magnitude, so that the natural crack detection result can be optimal. The excitation current was chosen to be 480A, the maximum current of the power generator used in the laboratory. Changing the excitation time affects the temperature distribution and heating efficiency, and increasing the excitation time does not contribute much to the detection, because the optimal observation time is just at the beginning of the excitation, so the excitation time is selected to be 300 ms.

The thermal imager is an SC655 thermal infrared imager produced by FLIR company, the equipment adopts a non-refrigeration type imaging sensor, the resolution of the equipment is 640 multiplied by 120, the frame frequency is 200Hz at most, and the temperature sensitivity is less than or equal to 50 mk. The electromagnetic induction excitation source adopts EASYHEAT0224 produced by Ameritherm in America, the power of the electromagnetic induction excitation source is 2.4kw, the maximum current is 480A, and the excitation frequency range is 150-. Two paths of pulses generated by the trigger can simultaneously control the thermal imager and the power generator. The cooling device adopts water cooling, the parameters are 240L/h of flow and 1.6kW of power.

In the experiment, firstly, a straight conductor of the inductor is selected, and the relative position of the straight conductor and a test piece is set; then turning on the PC, the water cooling and power generator, and paying attention to whether the straight wire is screwed tightly or not, so that water cannot leak; setting excitation current and excitation time according to experimental conditions; setting a file storage position, a trigger triggering mode, a thermal imager frame frequency and a thermal imager recording time on thermal imager upper software; and starting excitation, and recording the temperature information of the surface of the test piece.

(2) Description of weld seam Natural crack test piece

The tested piece is a pressure vessel welding seam. The stress accumulated on the near surface of the welding seam of special equipment such as a pressure container and the like can form tiny natural cracks, the width of the natural cracks can be as small as micron magnitude, the number of the natural cracks is unknown, the natural cracks are concealed and dangerous, and the tiny cracks are easy to accelerate to grow and cause accidents due to the fact that the natural cracks are influenced by mechanical load and environment for a long time in the service process. It is very significant to detect it. Shallow surface layer microgrooves have a closed characteristic, an unknown angle, clustering occurs, and gaps have substance filling, so that the detection sensitivity of electromagnetic thermal imaging is challenged.

Fig. 19 is a model diagram of a weld, a sector area is a weld area, the remaining height of the weld is 2mm, and a lift-off h in this experiment refers to the distance from a straight wire to a base material. According to the model diagram, the lift-off of the welding material part is different from that of the base material part, the eddy current density obtained by induction is different, and the detection difficulty is increased by the lines of the welding wave on the surface of the welding seam.

A schematic of the defects in the weld is shown in fig. 20. Fig. 21 is a permeation test result of a defect, and since the permeation test result magnifies and displays an actual defect, the actual defect is almost invisible to the naked eye. The photograph of defect 1 and the results of 60-fold microscopic magnification are shown in FIG. 22.

The size of the real defect is very small, and the requirement on the uniformity of the eddy current field in detection is very high. In addition, from the amplification result of the defects, the defects are discontinuous and have internal adhesion, so that the conductivity at the defects is not strictly 0, which also increases the difficulty of defect detection. Meanwhile, the thermal mode of the defect and the artificial crack are different, for example, the artificial crack lights up at the tip of the defect, and there are a hot zone and a cold zone. However, discontinuous natural cracks have a high electrical conductivity at their points of continuity and at internal bonds, and correspondingly, hot spots may also occur.

The experiment that the electromagnetic thermal imaging detected has been carried out to welding seam natural crack under different lift-off distance and different horizontal distance to this embodiment to verify that interval eddy current theory is effective to natural crack detection, and seek the method that improves natural crack detection effect.

(3) Experiment of different lifting heights of natural cracks of welding line

In order to verify the influence of different lift-off distances on natural crack detection, a straight wire is placed right above a defect, and experiments under different lift-off distances are performed on the defect 1, the defect 2 and the defect 3.

a. Experiments at different lift-off distances for Defect 1 and Defect 2

The results of the experiments for

defects

1 and 2 at different lift-off distances (2mm, 4mm, 6mm, 7mm, 8mm, 10mm) are shown in FIG. 23. Unlike an artificial crack that lights up at the tip of a defect, both defects are lit up over the entire area. In fig. 23(a), the lift-off distance is too small, the defect signal is weak, and is buried in noise; in fig. 23(b) and (c), the defect 1 and the defect 2 are partially submerged by noise; in FIG. 23(d), the defect signal is strong and can be separated from the noise; in fig. 23(e) and (f), both the defect signal and the noise signal are weak. Therefore, in this set of experiments, the optimum lift-off distance was 7 mm.

The black dashed area in fig. 23 is a high temperature area generated on the test piece by the excitation of the straight wire, and due to the existence of the weld reinforcement, the lifting-off of the welding material and the base material is different. In the present embodiment, the lift-off refers to the lift-off of the wire from the base material, and when the lift-off is small, the high temperature regions at the welding material and the base material are not aligned (fig. 23(c)), and when the lift-off is further increased, the influence of the residual height is reduced, and the high temperature regions at the welding material and the base material are aligned (fig. 23(d) (e) (f)). The defect signal in fig. 23(c) is covered by the high temperature area on the solder material and is difficult to extract from the strong noise signal, while the defect signal in fig. 23(d) is outside the high temperature area on the solder material and is separated from the noise signal. In this experiment, the key to detecting defects is to leave the band-shaped high temperature region uncovered with the defect signal.

In the process of increasing the lift-off distance, when the lift-off distance is 2mm, the straight lead is tightly attached to the surface of the welding seam, a high-temperature area is generated on the surface of the welding seam, and the eddy current distribution is not uniform enough, so that the disturbance of the defect is not better than the change of the self boundary of the eddy current field generated by the straight lead, and the defect is almost submerged; when the lift-off is 4mm and 6mm, the vortex distribution is still not uniform enough,

the defect temperature signal is swamped; when the lift-off is 7mm, the eddy current distribution is more uniform, the strength of the defect temperature signal is increased, and the defect temperature signal can be separated from the noise signal with equivalent strength; at 8mm and 10mm lift-off, the uniformity of the eddy current distribution is better, the noise signal of the weld surface is very weak, but the defect temperature signal is also not weak enough because of the eddy current density.

TABLE 4

Table 4 thermal contrast at different lift-off distances for defect 1 and defect 2. The results of the thermal contrast calculations for this set of experiments are shown in table 4, and from table 4 it can be seen that the thermal contrast of defect 1 and defect 2 generally increases and then decreases during the increase in lift-off distance, reaching a maximum at a lift-off of 7 mm. This is consistent with the analysis of the experimental heatmaps described above. The best lift-off distance for this set of experiments was 7 mm.

b. Defect 3 experiment at different lift-off distances

The results of the experiments for defect 3 at different lift-off distances (2mm, 4mm, 6mm, 9mm) are shown in FIG. 24. In fig. 24(a), defect 3 is mostly buried in the banded region created by the straight wire, only one end of the defect can be seen; in FIG. 24(b), the straight wire does not produce a band-like region on the test piece as in FIG. 24(a), but only the weld surface has an elliptical high temperature region, and the defect 3 can be clearly seen; in fig. 24(c), the high temperature region of the weld surface is circular in shape, the signal is strong, and defect 3 can be detected, but the signal is weak; in FIG. 24(d), since the noise signal at the weld surface is too strong, the signal of defect 3 is covered, and only one hot spot on defect 3 can be seen.

That is, when the lift-off distance is increased from 2mm to 9mm, first, the eddy current density is large and the defect 3 is submerged by a strong signal generated by a straight wire (fig. 24 (a)); then, the eddy current density is weakened to weaken the noise signal, the uniformity is improved to strengthen the defect signal, and the defect signal and the noise signal are equal in intensity, so that the defect can be clearly seen (fig. 24 (b)); finally, the uniformity is enhanced and the eddy current density is too weak, so that there is only a small disturbance around the defect, where the defect signal strength is weaker than the noise signal, and the defect can be weakly detected (fig. 24(c)) until it is substantially undetectable (fig. 24 (d)).

Distance of lift-off	h＝2mm	h＝4mm	h＝6mm	h＝9mm
					Thermal contrast	0.0087	0.0968	0.0426	0.0035

TABLE 5

Table 5 thermal contrast at different lift-off distances for defect 3. The results of the thermal contrast calculations for this set of experiments are shown in table 5. The thermal contrast was low at both lift-off of 2mm and lift-off of 9 mm. The thermal contrast at 4mm lift-off was 0.0968 and the thermal contrast at 6mm lift-off was 0.0426. Therefore, the optimum lift-off distance in this set of experiments was 4 mm.

(4) Experiment of natural crack of welding seam at different horizontal distances

Similar to the experiment of the lifting distance, the lifting distance is set to be 2mm, namely, the straight lead is tightly attached to the surface of the welding seam, and the experiment under different horizontal distances is carried out on the defect 1 and the defect 4.

a. Experiment at different horizontal distances of Defect 1

The experimental results for defect 1 at different horizontal distances (2.5mm, 3.5mm, 4.5mm, 5.5mm, 6.5mm) are shown in fig. 25. In fig. 25(a) and (b), the eddy current density is large, the temperature signal near the straight wire is strong, and the temperature signal of the defect 1 is weak due to poor uniformity, and thus the defect 1 is partially submerged; in fig. 25(c) and (d), the decrease in eddy current density attenuates the temperature signal near the straight wire, while the increase in eddy current distribution uniformity enhances the temperature signal at defect 1, which is equivalent to defect 1, and defect 1 can be seen; in fig. 25(e), the value of the eddy current density is too small, and the temperature signal generated by the straight wire is hardly observed in the vicinity of the defect 1, but the temperature signal of the defect 1 is too weak to be detected.

Horizontal distance	m＝2.5mm	m＝3.5mm	m＝4.5mm	m＝5.5mm	m＝6.5mm
						Thermal contrast	0.0172	0.0390	0.0996	0.1106	0.0070

TABLE 6

Table 6 thermal contrast at different horizontal distances of defect 1. The results of the thermal contrast calculations for this set of experiments are shown in table 6. From the thermal contrast, the thermal contrast is lowest at a horizontal distance of 6.5 mm. The thermal contrast is slightly greater at horizontal distances of 2.5mm and 3.5 mm. The thermal contrast was higher for horizontal distances of 4.5mm and 5.5mm, with the highest thermal contrast at 5.5 mm. The optimal horizontal distance for this set of experiments was 5.5 mm.

b. Experiment at different horizontal distances of defect 4

The results of the experiments for defect 4 at different horizontal distances (3.5mm, 4.5mm, 5.5mm, 6.5mm) are shown in fig. 26. The white dotted line in the figure indicates the rectangular area where the defect 4 is located, the part of the defect 4 close to the straight wire is buried, and the tip far from the straight wire is marked by a white arrow. When the horizontal distance is too small, the tip is buried (fig. 26(a)) or cannot be completely separated from the noise signal (fig. 26(b)), and when the horizontal distance is too large, the defect signal is too weak to be detected (fig. 26(d)), and the tip can be seen only in fig. 26(c) where the horizontal distance is moderate. Since only the defect tips were visible in this set of experimental heat maps, the thermal contrast was no longer calculated. The optimal horizontal distance for this set of experiments was also 5.5 mm.

(5) Analysis of test results of natural cracks of welding seams

The above experiment results show that the experiment of the natural crack of the welding seam is consistent with the simulation result of the natural crack under different lifting distances and different horizontal distances, and both the experiment result and the simulation result meet the interval eddy current effect. Along with the increase of the lifting distance and the horizontal distance, the density of eddy current on the surface of the test piece is weakened, and the uniformity is improved. Therefore, the surface of the test piece has three areas, and the area A closest to the straight conducting wire has the strongest eddy current density but poorer uniformity; the eddy current density in zone C, furthest from the straight wire, is too low to detect defects; the middle B area has enough eddy current density and uniform eddy current field. Since natural cracks have high requirements on the uniformity of detection, the selection of an optimal detection area is crucial.

TABLE 7

Table 7 shows the optimum detection distances in the simulation and experiment results in this example. A horizontal distance of 0 indicates that the straight wire is directly above the defect. As can be seen from table 7, the values of the optimum lift-off distance obtained by the fixed horizontal distance and the optimum horizontal distance determined by the fixed lift-off distance are not fixed under different experimental conditions.

The magnetic permeability of the material, the magnitude of the excitation current and the size of the defect are all factors influencing the optimal lift-off and distance. The magnetic permeability of the material and the size of the exciting current influence the eddy current distribution, and the size of the defect influences the disturbance size of the tip of the defect, so the requirements on uniformity are different. Therefore, the determination of the detection area is a more complicated problem.

Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.

Claims

1. A pulse eddy current thermal imaging defect detection method suitable for natural cracks is characterized by comprising the following steps:

(3) defining an evaluation index: thermal contrast Δ T

2. The method for detecting the defects of the natural cracks by the pulse eddy current thermal imaging according to claim 1, wherein the straight wire lift-off distance and the horizontal distance are determined according to the principle that the thermal contrast Δ T is maximum as follows: fixing the lifting distance of the straight wire, changing the horizontal distance, determining the horizontal distance or fixing the horizontal distance of the straight wire when the thermal contrast delta T is maximum, changing the lifting distance, and determining the lifting distance when the thermal contrast delta T is maximum.